Twitter page for Data Ethics and Governance Section of the Royal Statistical Society. Views expressed are authors' and not necessarily official RSS policy.
When the Code of Practice for Statistics was formulated, production focused on regular publications and direct data collection, but is it suitable for the contemporary professional environment? The practices need to cover all the issues that can be a concern, even if unrealised
Many pernicious practices have proliferated, so it is only a first step to set out what they are and how they work against the interest of consumers. We look forward to seeing how the recommendations of this position paper have an impact on online environments
A lot has been said about data at the pandemic public inquiry. But most is to point at the problems, without really setting out the designs for solutions for each. We hope @RoyalStatSoc will be asked to contribute to further modules
On behalf of the RSS, our president @AndrewDGarrett has responded to module one of the @covidinquiryuk, which focuses on pandemic preparedness
Read our 12 recommendations for the UK’s statistical and data ecosystem ⬇️
https://t.co/GvZae94tHw
@RoyalStatSoc Current practices are framing AI systems as taking on narrowly defined interactions with a clear precedent, including simple binary decisions which are evaluable against an existing standard, but that is not where the potential impact (risk or benefit) of AI systems lies
@RoyalStatSoc One concern is around having the methodology to evaluate the impact of AI systems on society, especially as these systems interact over sustained periods of time, and automate human processes
In all the noise about technology, the UK (& Canada) has been quietly leading the world in data ethics. Not corporate statements, not flashy government announcements, nor civil society campaigns. Pragmatic government statisticians @UKStatsAuth@StatCan_eng
https://t.co/H6OHdMeuHZ
Have people who work with data in your organisation but don't know how to support their professional development? Find out how with the RSS via our corporate membership scheme: https://t.co/lSFmYBQqLC
The @CommonsPACAC is inquiring on 'Transforming the UK evidence base', including asking "What does it mean to use data ethically, in the context of statistics and analysis?"
https://t.co/i6DDWgTQah
The @RoyalStatSoc was among those sending a response to the AI regulation whitepaper consultation last week, and apparently "it has received a range of highly informed critiques, and praise from several surprising quarters"
https://t.co/A1ZhtqjZ96
Nice update from the (world leading) @UKStatsAuth Centre for Applied Data Ethics - in prospect is further training material for analysts and a further evolution of their ground breaking self-assessment tool
https://t.co/TBgWj4ZlUq
There is more to safeguard than the business models of large technology companies - we are encouraging the @RoyalStatSoc
to speak for professionals as other leadership is most disappointing https://t.co/38TIuN5nhh
The @RoyalStatSoc has developed the criteria for honorary fellowships to require commitment and support for the values of the statistical community, beyond recognition for exemplary contribution. Nominations are open now, until the end of October:
https://t.co/ChDDgdxXga
@StatsRegulation@lrcfs @unecestat We are thinking about other angles for this discussion, including tying together some of the threads drawn out in these discussions. If you have a view on what else to cover, let us know
There are three sessions recorded now discussing the misuse of statistics with @StatsRegulation - our session in London, in March of this year with more practical issues seen by professionals:
https://t.co/9x1yZZF2XM
@StatsRegulation@lrcfs And a more international discussion about regulation, from the celbration of 30 years of the @UNECEstat Fundamental Principles of Official Statistics, in March of last year
https://t.co/pTdr118FlO
The boards of far too many organisations rely on dashboards when their responsibility is to understand the assumptions about risk embodied, often implicitly in the models supporting the data. We believe the majority lack the capability (and will) to do better
Errors in government models can cause issues for critical public programmes.
Our guide for government and agency audit committees will help you question the quality assurance of models and assess if model risks are effectively managed.
Read our guide 👇https://t.co/e7mkdDOJoD
📣We are inviting submissions of discussion papers on the analysis of citizen science data
Papers selected will be presented at a discussion meeting at our 2024 conference & published in an RSS journal
More details & submit your abstract by 30 June ⬇️
https://t.co/Vb45SxqNod